Slashdot Mirror
Ogg Vorbis Changes (Just About) Everything
The good people of Ogg Vorbis have a new beta release out (number 4) for which they claim better compression, nicer sound, fewer bugs and more protein than the last. While that's nice enough, that's not the only news on the Vorbis front: probably more important in the long run is that the guys behind Vorbis have formed a non-profit called Xiph.org to replace the S-class corporation they've been developing as for a while, Xiphophorous. Emmett of BinaryFreedom had
a cool chat with Vorbis developers Christopher Montgomery
and Jack Moffit about the new release,foundation, encoding, and hardware capable of playing back the Vorbis format -- well worth reading. Plus, you can
download the new beta (and some sample tunes), too. Oh, yes, and there's the little matter of moving from the GPL [?] to BSD license [?] , with what they say is RMS' blessing. You will have to read to find out why, though;)
Is it just me or is that Jesus hitting a snake with an ax? What the heck does the logo mean?
Well at any rate, I have used the betas of the past from Ogg Vorbis and they work quite well. Keep up the good work guys.
"I am sorry, I switched to a new ISP becuase you guys dont offer Yahoo like AOL does!" - annonymous customer
This is actually not surprising, he had a similar reasoning for making the gzip compression code available on a BSD-like license.
It will probably be too much to hope for, that some of the "RMS will only accept GPL" people will take note.
I'm conflicted about this: on the one hand, I am concerned that companies will glom on to Vorbis, make proprietary extensions, and not release them back into the free software pool. Not good.
On the other hand, as a professional embedded software developer, I have a need that Vorbis would be just perfect for. Under a BSD license, I would have no problems with using it (due to constraints beyond my control, the code would have to be linked against some decidedly CLOSED SOURCE code, thus chucking the GPL out the window). However, I was perfectly willing to go to my managers and have them negotiate a license with Xiph to allow use to use the Vorbis code under a closed-source license and pay them money for the privilege (while maintaining the normal license as GPL). That will be a great deal harder to justify now....
www.eFax.com are spammers
If you can get to the heavily slashdotted interview on binaryfreedom.com, you'll see RMS's comments on the license change.
Why do I get the feeling that if you asked RMS's opinion on slaughtering the innocent for the glory of Satan, and bathing in their warmly splashing blood, he'd reply I'm all for it, as long as you make it clear I support "Free Software" and not "Open Source", and don't imply I agree that there is such a thing as a "Linux operating system".
But seriously though.. the reason the Xiph folks gave for their license switch is that they want Ogg Vorbis to be "a basic building block of the internet for multimedia", and hence chose to go BSD: minimise the restrictions on the source, maximise the chance that it will be widely adopted. Fair call - you gotta look at what you're trying to achieve, and ask yourself if every man and his dog stick my code into a proprietary app, does that help my cause or hurt it? In this case, I think it's quite clear that it would help Ogg Vorbis if, to pick and example at random, Microsoft stuck a closed-source .OGG replay codec into Windows.
If quality is of any importance to you then you don't want to convert from MP3 to Vorbis, as they're both lossy with their compression and you're very likely to start losing important data going from one to the other. It's the same with any lossy formats, any data type.
You'd be better off keeping the MP3s you have and just doing new ones with Vorbis, unless you really do have enough time to re-rip them.
Last time OV was mentioned on /. (the last beta, presumably) I download the same encoder and xmms plugin for playback. I encoded a couple of CDs and tried it out. Here's what I found:
First, the sample encoder is MUCH easier to use than what I've already been using (GRip). I don't know if that's because my current method is so terrible or because the new one is so great.
Second, the resulting files were about 10% smaller. Others may say "so what, hard drives are cheap", but:
1) I only have 4.5 GB and don't have the extra cash to buy larger.
2) Larger hard drives make a 10% savings even MORE worthwhile. Consider: If I saved 10% of a 4 GB drive, that's 40 MB--room for maybe 10 additional songs or about one CD. But if I saved 10% of a 400 GB drive, that's an extra 4 GB--enough for 100 CD's.
Third, the sound quality was "equivalent". That is, I couldn't tell the difference, BUT I'm not an expert and my sound equipment is FAR from top of the line (just some computer speakers plugged into an AWE32).
--
Non-meta-modded "Overrated" mods are killing Slashdot
Non-meta-modded "Overrated" mods are killing Slashdot
(Hey Ryan! Here's your proof!)
Using slightly different compression methods, how does OGG performance compare to MP3? Systems being as fast as they are, nowadays, this isn't that *big* of an issue, but it is nonetheless with older systems in mind and those with heavy load.
In my experience, an Ogg Vorbis file compressed with variable bit rate centred on 160kb with minimum 128 and maximum 192 is the same size or smaller than a 128 kb MP3 file compressed from the same source CD. The quality of the encoding is excellent - I'm extremely impressed with the fidelity and sound of the Ogg Vorbis output, and I've now standardized on the Ogg Vorbis encoder for all my on-disk music.
Currently the encoder I'm using (the beta3 release under CDex on my WinNT box and under Grip on my Linux box) encodes at about 1xCD data rate on a 400MHz PII machine. Thats not as fast as some of the MP3 Encoders - the x86 assembler ones can acheive around 2.5x on the same machine. But it is early days for the Vorbis encoder - I expect it to get a lot faster once the 1.0 release is out.
For those who say OGG is late, consider the factors in it not being so pushed for. There was never a huge consumer demand for an MP3 alternative. People own gigs and gigs of MP3s... telling them to convert because of a patent that will affect them when they purchase a commercial product by a few dollars doesn't mean much to them, as they commonly use only XMMS/Winamp and Napster/Gnapster.
I don't see a lot of people immediately switching all of their collection to Ogg Vorbis. However, all the important players support Vorbis codecs now and therefore it is a snap to start adding .ogg files into the collection. I am slowly replacing my remaining .mp3s with .ogg versions as I get time and anything new is automatically encoded as .ogg.
Companies looking to market commercial digital music players and/or software, on the other hand, plagued with the prospect of paying the MP3 patent owner money for each product they sell, must be more interested. But, again, it is very dependent on the consumer since they would have to convert the MP3 to OGG without help from any software supplied by the commercial company supplying the product -- but software supplied from a non-commercial entity, such as Ogg Vorbis creators, could be downloaded.... Packaged, though? I don't think it could be packaged with the product, regardless of it being "free" or not, because it would be included as part of a commercial product. The best a company could do would be an automatic download (of course with yes/no prompt and license agreement) of the extension from Vorbis to their uploading software.
Ah - but here the use of a BSD license is important. Because the Commercial packages can compile the Vorbis support into their own systems and not be forced to distribute the source code, modified or otherwise, they have no real reason to neglect this format - they can add it for free and add yet another feature to those marketing tick-lists. Like I said, conversion from MP3 to OGG is probably a non-starter but once this format starts to attract the attention it deserves, I would expect it to become fairly popular fairly quickly.
Cheers,
Toby Haynes
Anything I post is strictly my own thoughts and doesn't necessarily have anything to do with the opinions of IBM.
there would be no Eminem, no N-Sync and no Limp Biskit.*
;)
;)
And this is a bad thing?
We are seeing more and more the danger that MP3s present to musicians' livelihoods
Just so people understand: Digital Music Distribution only threaten the lively hood of the RIAA. Music (and the Arts) thrived before IP law. The Arts will thrive long after people stop trying to impose artificial constraints on culture/thought/speech (IP law). Please read this article by Courtney Love. Members of the band Garbage are caught up in a legal fight with RIAA thugs. What was the name of that black singer who went bankrupt a couple years ago - she had sold millions and millions of records - and was penniless? The RIAA is messing with public perception - digital music distribution threatens them (and they are obsolete but are _BUYING_ laws to protect their pocket book schemes) and not artists.*
*These people are essentially *not* artists - they are products. This is what the RIAA is interested in; not Art. _BUT_ unfortunately i cannot tell *other* people what to listen too - just as I cannot convince 95% of the worlds sheeple how to vote, or why it is they think they like these 'artists' (because the public is horribly connected to The Media Machine(TM) which purpose is to end individual thought and replace it with a 'shared' group experience which centers on the Consumer Values(TM) - I believe 95% of people are incapable of making decisions based on objective opinions because they are caught up in a massive experiment in Population Control (no I don't mean there is a dark force controlling it all - but the 'marketing' machine evolves to serve those who intend to serve themselves - which is why the system is becoming effective - its goals are to homogenize thought to match product offerings.. and its working.) Hows that for paranoid?
Actually, that sentence was out of a longer letter that was an RMS reply to Jack. There was some cut-and-thrust debate going on, and RMS's quote was a sensible reply to an earlier assertion made by Jack. The whole "Linux operating system" thing didn't just come out of the blue, the interview just ended up with it edited in a somewhat unfortunate sounding way for RMS ;-)
Monty
xiph.org
The logo is sort of obscure, but the snake/sine wave thing is fairly obvious, and everybody likes to see powerful mythic figures hitting things with hammers.
But what in the WORLD were they thinking with the name? Ogg Vorbis? Nowhere near the "catchiness"
of saying mp3. Not to mention that any format must have a great three-letter acronym to catch on. I think "xiph" is a great name for the format, and XPH would make a catchy TLA.
Please guys, change the name, or adopt such a TLA. The name "Ogg Vorbis" just sounds way too plan-9-from-outer-space geeky.
--
Libertarianism is rich wolves and poor sheep playing gambler's ruin for dinner.
For Stallman, any other form of licence is just a tactical compromise on the way to finagling everyone into using his beloved GPL. He wants to try to promote Ogg to become the de facto standard, and then start including features in it with GPL code, so that anyone who wants to stay up with the development path has to join his merry band of intellectual property guerrillas. Of course, when Microsoft do this, it's called "embrace and extend", but Open Source's favourite sweaty hippie would never do anything so bad, wold he?
Look, that makes no sense. Just sit back and think about it. Microsoft "embrace and extend" = create incompatable versions so that everybody is locked into using Microsoft products/standards, and so that Microsoft gets sole control over something which was once out there for all to use. GPL "embrace and extend" = forever make it impossible for any one self-interest to use something to their own ends without leaveing it out in the open for anybody else to use.
How is this comparable? One is inherently selfish. The other is inherently protective. They're opposite.
Sure, criticize the restrictiveness of the GPL. That's fine. But that restrictiveness is of a *very* different nature than the restrictiveness of proprietary licenses such as what comes out of Microsoft, and it's just a stupid troll to try to compare the two.
Regarding the "IP guerrilla" nature of RMS and the GPL: sure, weakening IP is their goal. On the other hand, their way of going about it is entirely fair, and calling them a guerrilla isn't. They aren't going in there and insisting that proprietary software be made illegal, or that proprietary software must be opened up. They are *suggesting* that you might want to choose not to use it. What the GPL does is insure that that which *starts* open, *stays* open. What's so awful about that? It sounds like a damn good idea to me. If you're going to defend propreitary software, bear in mind that almost nobody who produces such software would ever let anybody else use their code without all sorts of restrictive licensing terms dictated by *them*. These restrictive licensing terms will tend to be must less protective of "general use" than the GPL is.
-Rob
In beta 2, oggenc's default encoding mode was 160kbps. In beta 4, the default is 128kbps. There *is* an audible difference, even with the improvements to 128kbps since earlier betas.
Monty
xiph.org
This is a perfect example of someone choosing a license based on what their goals are for the code, rather than religious beliefs or whatever. IMO all of the BSDL vs. GPL debates are pretty pointless, since most of the time the participants are arguing that one or the other license is ideal for all open source software. I say, it all comes down to what your goals are for the code you're writing; pick whichever license is most appropriate, rather than mindlessly advocating using one over the other for everything. This goes for proprietary software too.
Does anybody know about support for multichannel (specifically 5.1) audio streams? I've got a surround setup and some surround recordings on DVDs, and it would be just swell if I could encode those...
Vidi, Vici, Veni
Every Ogg + XMMS crash bug we know about is fixed... in XMMS CVS. Because there have been no major releases of XMMS recently, most people are still downloading and using buggy 'official' XMMS packages. Thus, folks are unwittingly reporting bugs that have been fixed for weeks or months. A number of the crash bugs were in the Ogg plugin, but a number have also been in XMMS itself, which Ogg simply had the bad luck to tickle. Upgrading the plugin alone can't save you.
So, first grab and build XMMS from XMMS CVS. It actually builds cleanly with minimum fuss. At that point, if you get a crash playing Ogg, the XMMS developers and we would very much like to hear about it.
Monty
xiph.org